Pairing Two Gut Hormone Drugs Protected Mice From Immune Injury

A drug that did nothing by itself to shield mice's intestines from a destructive immune attack worked once researchers paired it with a second hormone-mimicking drug built for a completely unrelated disease.

The finding, published October 20 in Cell Reports Medicine by a University of Toronto team led by longtime GLP-1 researcher Daniel Drucker, is the fourth paper in just over six years from overlapping research groups chasing the same question with mixed results: can two gut hormones, GLP-1 and GLP-2, help the intestine survive a T-cell attack, the kind of immune damage seen in conditions like graft-versus-host disease after a stem cell transplant?

Two hormones, two unrelated drugs

GLP-1 and GLP-2 are made together by the same intestinal cells, called enteroendocrine L cells, in response to a meal. GLP-1 slows digestion and prompts insulin release — the biology behind Novo Nordisk's semaglutide, approved by the FDA for type 2 diabetes and, at a higher dose, for chronic weight management. GLP-2 does something different: it signals the intestinal lining to grow and repair itself, which is why Takeda's teduglutide, a lab-made version of it, is approved instead for short bowel syndrome, where patients depend on intravenous nutrition because they don't have enough functioning gut left.

Because GLP-2 signaling is about gut repair, researchers have spent several years asking whether boosting it — or boosting GLP-1 alongside it — could also help the intestine survive a very different kind of damage: an attack by the body's own T cells, newly transplanted or newly activated, on the gut lining itself.

A rocky run of mouse data

The case for teduglutide looked strong in 2020. A team reported in Blood that the drug cut rates of acute and steroid-resistant graft-versus-host disease across several mouse models, apparently by helping the gut's stem cells and Paneth cells regenerate. In biopsies from people who had received stem cell transplants, the same researchers found that fewer hormone-making L cells and higher blood levels of GLP-2 lined up with worse survival — a correlational hint that the biology mattered in humans too, though it fell short of a drug trial.

The Drucker lab then tried to build on that work and ran into trouble twice. In a paper posted online in August 2025 and published that October in Molecular Metabolism, the group reported that semaglutide alone did nothing: gut T cells did carry a working GLP-1 receptor, but dosing mice with semaglutide failed to lower cytokine levels, gut injury or disease severity in experimental graft-versus-host disease. A second paper from the same group, accepted that December and also in Molecular Metabolism, could not reproduce the original teduglutide benefit either, despite using what the authors describe as an identical protocol. One finding did hold up: mice engineered to lack the GLP-2 receptor altogether died faster and developed more bacteremia, so the hormone's own natural signaling still seemed to matter for gut defense — the drug boost just hadn't shown up as protective in their hands.

That's where this month's paper comes in. Working with mice genetically missing GLP-1 receptors, GLP-2 receptors, or the single gut gene that makes both hormones, the team activated the animals' T cells and found the hormone-deficient mice suffered worse intestinal damage than normal mice. They then treated the hormone-deficient animals with teduglutide, semaglutide, or both together — and the damage eased in all three treated groups.

What's established, and what's still open

A few things are worth weighing before reading too much into the newest result. Every study in this chain, including the latest one, is in mice; none of it has been tested as a treatment in people with graft-versus-host disease or any other T-cell-driven gut condition. The injury model in the new paper is a general T-cell-activation setup in genetically hormone-deficient animals, not necessarily the same full transplant-rejection model the 2020 and 2025 papers used, so "it worked this time" and "it didn't work last time" aren't a clean head-to-head comparison. The positive results so far also come from one overlapping set of labs; independent replication by outside groups hasn't happened yet. And the 2025 failure to reproduce the original 2020 teduglutide effect is itself a useful reminder that single mouse studies, even carefully designed ones, rarely settle a question on their own — protocol details that look minor on paper can change an outcome entirely.

What this means for researchers

The throughline across four papers in six years is that the body's own GLP-1 and GLP-2 signaling appears to matter for how the gut survives a T-cell attack, even when giving the matching drug by itself hasn't reliably reproduced a benefit. That gap between "the natural hormone matters" and "the drug version of it works" turns up often in peptide pharmacology, and it's exactly what a null or mixed result is supposed to flag rather than quietly bury.

It also suggests that testing a single hormone pathway against immune-mediated gut injury may be asking an incomplete question. An earlier post on this journal covered a trial where adding a third hormone to an existing GLP-1 drug pushed weight loss well past what either component managed alone — different biology, but the same pattern of combinations outperforming solo treatment. If the new gut-injury result holds up, the variable worth isolating isn't semaglutide or teduglutide on its own, but whether disrupting both the metabolic and the repair pathway together is what exposes the intestine, and whether restoring both at once is what's needed to protect it.

None of this says anything yet about dosing, timing, or which patients might benefit — this combination has reached mice, not clinics, and the researchers themselves will need to repeat the experiment in the classic transplant model the earlier papers used before anyone can say the discrepancy is resolved rather than just moved. Two drugs built for entirely unrelated diseases ended up sharing a cage of mice; whether they'll ever share a patient is a much longer way off.